Many nonproteogenic amino acids have proved useful for inhibiting biodegradation and improving biological activity in peptides and peptidomimetic drugs. Few non-genetically encoded branched-chain amino acids (BCAAs) are commercially available, despite the importance of BCAA side-chain interactions in determining polypeptide structure. Chiral branches permit fine-tuning of biological activity by subtly changing side-chain shapes. For example, D-isoleucine substitution gives more specific and effective insulin and vasopressin antagonists, and potent short antiangiogenic peptides. Branch-carbon configurations of beta-methyl arylamino acids strongly affect activity. Thus beta-chiral BCAAs can provide better models for bioactive polypeptide conformations and greatly improve both activity and duration of action in peptide therapeutics. Most syntheses of beta-chiral BCAAs begin by stereorandomly building carbon skeletons, then separating diastereomers and finally enantiomers. In the case of D-alloisoleucine, numerous attempts to improve on this inefficient synthetic strategy have only resulted in expensive, complex processes which are difficult to scale up. Interest in less-common branched-chain amino acids is high, but commercial sources are currently not providing the quantities needed for drug development at acceptable cost. In developing a scalable enzymatic process to cleanly isomerize L-Ile to D-allo-Ile, we realized that obtaining amino acid frameworks with the correct side-chain branch configuration is the crucial problem in making any beta-branched BCAA, because stereo directed epimerizations can quantitatively convert alpha-isomeric mixtures to homochiral )roducts. We will compare the synthetic and economic merits of straightforward glycine anion alkylations with two novel cyclopropane ring-opening procedures for making amino acids with beta-chiral branches. alpha-Carbon epimers will then be made uniformly D- or L- by well-precedented chemoenzymatic processes.